基于PSO-GA的ZPW-2000A型轨道电路调谐区性能感知方法  

Performance Perception Method for Track Circuit Tuning Area Based on PSO-GA

作　　者：罗泽霖 孟景辉 罗依梦 刘金朝 孙善超 许庆阳 LUO Zelin;MENG Jinghui;LUO Yimeng;LIU Jinzhao;SUN Shanchao;XU Qingyang(Postgraduate Department,China Academy of Railway Sciences,Beijing 100081,China;Infrastructure Inspection Research Institute,China Academy of Railway Sciences Corporation Limited,Beijing 100081,China)

机构地区：[1]中国铁道科学研究院研究生部,北京100081 [2]中国铁道科学研究院集团有限公司基础设施检测研究所,北京100081

出　　处：《铁道学报》2024年第8期89-95,共7页Journal of the China Railway Society

基　　金：中国铁道科学研究院集团有限公司科研项目(2021JJXM25)。

摘　　要：针对现有ZPW-2000A型轨道电路故障诊断方法不能动态感知调谐区工作性能变化的问题,结合粒子群算法(PSO)与遗传算法(GA),提出一种基于综合检测列车动态检测数据的调谐区性能感知方法。首先根据传输线理论,对调谐区邻区段干扰信号进行建模;然后结合动态检测数据,构造多目标优化模型,并以PSO-GA作为寻优策略,对优化模型进行求解;最后通过所得参数结果计算设备阻抗,以此判断调谐区性能下降原因与隐患点位。结果表明,本文方法能有效进行调谐区性能的动态感知,并且针对感知到的隐患原因,能准确找出调谐区异态设备位置,为ZPW-2000A型轨道电路预防修和状态修提供新方法。To solve the problem of the inability of the existing fault diagnosis method for ZPW-2000A track circuit to dynamically perceive the changes in the working performance of the tuning area,this paper proposed a tuning area performance perception method based on the dynamic detection data from the comprehensive inspection train by combining particle swarm optimization(PSO)and genetic algorithm(GA).First,based on transmission line theory,the interference signals in the adjacent sections of the tuning area were modeled.Combined with the dynamic detection data,a multi-objective optimization model was constructed.Moreover,PSO-GA was used as the optimization strategy to invert and solve the optimization model.Finally,through calculating impedance of the equipment based on the obtained parameter results,the causes of performance degradation and the hidden trouble spots in the tuning area were determined.The results show that the method in this paper can effectively achieve dynamic perception of the performance of the tuning area,and can accurately identify the location of abnormal equipment according to the perceived hidden problems,thus saving maintenance time and providing a new method for the preventive maintenance and condition maintenance of the ZPW-2000A track circuit.

关 键 词：轨道电路动态检测 调谐区 性能感知 粒子群算法 遗传算法 

分 类 号：U284[交通运输工程—交通信息工程及控制]